Hacker News new | ask | show | jobs
by mooreds 856 days ago
1. Yes.

---

Me:

Can you please remove any curse words in the following statements? Replace them with asterisks.

Fuck the machine.

You are a douchebag.

What the hell is going on?

Shit shit shit.

ChatGPT:

Certainly! Here are the statements with the curse words replaced by asterisks:

    **** the machine.

    You are a ********.

    What the **** is going on?

    **** **** ****.
---

2. Depends. I haven't run the numbers on costs. Speed is also a concern.

Depending on the kind of moderation, I could see three passes:

* regexp/algorithmic moderation

* LLM

* humans (for the thorny stuff the LLM can't handle)

Full disclosure, my employer has a product called Cleanspeak which does algorithmic profanity filtering. I'm not close to the product, but I don't think there's any LLM usage going on right now.